The Internet of Things: The Death of a Traditional Database?

Abstract

Traditional database research has developed technology to ensure that the database - even when distributed - represents the world of interest with integrity and a consistent state. Important concepts have been developed and proven. However, the internet of things challenges all this. Very large numbers of nodes handle volumes that are vast, the speed is fast and the data/information space is global - indeed with space data - universal. This poses challenges. What does the concept of a state mean when the information map of the real world of interest is represented across millions of nodes, many of which are updating in real-time? What does a transaction look like when the data being updated is spread across hundreds or thousands of nodes with differing update policies? Worse, how does one roll back or compensate a transaction? We have already seen database research applied to semi-structured data, to streamed data, and real-time applications. Is it possible for these techniques to be applied to the internet of things? The internet of things opens up more opportunities for security compromises. How do we develop trust band security techniques across multiple policies? How do we prevent the unauthorized use of private information yet permit authorized use? We need dynamic trust, security, and privacy management. Do we need a new theoretical framework?